L₁ adaptive control of a dual-rotor tail-sitter unmanned aerial vehicle with input constraints during hover flight

The input constraints of control surfaces may lead to saturations, which could limit the achievable performance or even cause instability of vertical take-off and landing (VTOL) tail-sitter unmanned aerial vehicle (UAV). To improve flight safety and attitude tracking performance, a novel $L-{1}$ adaptive control architecture for attitude tracking of a tail sitter subjected to input constraints is proposed in this study. The imprecise mathematical model, low weight, and small size all present different challenges when designing a control system. In this work, a feedforward compensator is first used to narrow down the uncertain bounds with the consideration of finite accuracy from prior knowledge. Second, according to the linearized model, a baseline controller is designed to offer basic performance for a nominal system without uncertainty. Finally, the $L-{1}$ adaptive controller is developed to compensate for unmatched uncertainties based on the control system developed before. The stability and performance bounds of the closed-loop system are analyzed to illustrate the impact of input constraints. The numerical simulations and flight tests are carried out to verify the improved attitude tracking performance when input saturation exists.